NN-PLS based on-line component soft-analyzer for Tennessee Eastman process

Filtered by the orthogonal signal correction technique, the selected process measurements projected by NN-PLS algorithm to build the soft component analyzer. Thus a new approach proposed to solve the delayed feedback quality signal of the Tennessee Eastman process by building this soft analyzer. As a result the quality indices can be obtained online in time with enough precision. The simulation results confirm that the soft-analyzer model has satisfying robust and predictive performance and more importantly this soft-analyzer could be as the basis of other advanced process control and on-line monitoring techniques.